Latest / Hardware Startups with Fexingo: Devices, Robotics, and Manufacturing Tech Companies / How Hardware Startups Use Potting Compounds for Protection
Transcript
- Lucas: So you've got a board full of sensitive components, maybe a sensor array or a power module, and you know conformal coating isn't going to cut it because the environment is too harsh — vibration, moisture, maybe even chemicals. That's when hardware startups start looking at potting compounds. Luna: Potting — that's where you basically encase the whole assembly in a solid resin, right? Lucas: Exactly. Instead of a thin film like conformal coating, potting fills every gap around the components with a thermoset plastic — epoxy, polyurethane, silicone. It becomes a solid block. The trade-off is that you lose repairability — once it's potted, you're not swapping out a chip — but you gain serious protection against shock, vibration, and moisture ingress. Luna: So when does a startup make that leap? Because potting sounds like a step up in complexity and cost. Lucas: It's usually when field failures are traced back to environmental stress. I talked to a drone startup last year — they were seeing about a 12 percent failure rate on their flight controllers after a few hundred hours of operation. The root cause? Micro-vibrations were cracking solder joints on a voltage regulator. They switched to a polyurethane potting compound with a shore hardness around D50 — not too brittle, not too soft — and failures dropped to under 2 percent. Luna: That's a huge improvement. And polyurethane specifically — why that over epoxy or silicone? Lucas: Great question. Epoxy is the most common — it's strong, adheres well, and has good chemical resistance. But it's also rigid and can generate a lot of heat during curing, which is called exothermic heat. If you pot a dense board with epoxy and it cures too fast, the internal temperature can hit 100 degrees Celsius or more, which can damage components. Polyurethane is more flexible and has lower exotherm, so it's better for vibration damping. Silicone is the most flexible and handles extreme temperatures — minus 50 to plus 200 degrees — but it's expensive and doesn't adhere as strongly, so moisture can creep along the interface. Luna: Right, so the material choice really depends on the specific failure mode you're trying to solve. Lucas: Exactly. And there's another dimension: thermal management. Potting compounds are generally poor conductors — most have a thermal conductivity around 0.2 to 0.3 watts per meter-kelvin. Air is about 0.025, so it's still better than air, but if you have a power component that generates a lot of heat, you might need a thermally conductive filler. Some epoxies come with alumina or boron nitride fillers that push conductivity up to 1 or even 2 watts per meter-kelvin. That can be a game-changer for LED drivers or motor controllers. Luna: But those filled compounds are thicker, right? Does that affect the dispensing process? Lucas: It does. Viscosity is a big factor. A standard unfilled epoxy might be like syrup, but a filled one can be like peanut butter. That means you need a different dispensing method — maybe a piston pump instead of a peristaltic pump, or you might need to preheat the resin to lower viscosity. Also, you have to worry about settling — the filler particles can sink to the bottom of the container over time, so you need to keep the material agitated. Luna: And what about air bubbles? I imagine trapped air can cause issues. Lucas: Air entrapment is one of the most common defects in potting. If you pour a high-viscosity compound over a board with tall components, air gets trapped under overhangs and in crevices. When the resin cures, those voids become weak points where moisture can condense or where stresses concentrate. Startups often solve this by vacuum degassing — they put the mixed resin in a vacuum chamber for a few minutes before dispensing, or they pot the assembly in a vacuum chamber so the air is pulled out as the resin flows in. Luna: So vacuum potting is a common technique for high-reliability applications. Lucas: Yes, and it's one of those things that sounds expensive but can be done with a relatively affordable vacuum chamber — maybe a few thousand dollars. For a startup doing low to medium volumes, that's a reasonable investment if it cuts failure rates significantly. Luna: And speaking of investment — I'm thinking about the listeners who are building hardware right now, maybe in a garage or a small lab. If this kind of deep-dive into manufacturing processes is helpful for what you're working on, we want to keep these conversations coming. Lucas: Yeah, and the way to do that is pretty straightforward. If today's episode was worth a coffee to you, you can find the link at buy me a coffee dot com slash fexingo. That's the entire business model — no ads, no sponsors, just listeners who find value and chip in. Luna: It really does make a difference. And back to potting — one thing I've seen debated is the question of repairability. Some startups design their boards so that critical components are left un-potted, or they use a silicone that can be cut open and re-sealed. Is that practical? Lucas: It's a compromise. Silicone is softer and can be cut with a knife, but resealing is messy and rarely as good as the original. Some companies use a 'conformal coating plus selective potting' approach — they coat the board with a thin conformal coating first, then pot only the high-stress areas like connectors or large capacitors. That way, the rest of the board remains serviceable. Luna: That seems like a smart middle ground for startups that are still iterating on their design. Lucas: Absolutely. And there's another technique called 'glob top' — where you dispense a small dot of epoxy over a single chip or a bond wire, like what you see on some SD cards or key fobs. That's not full potting, but it's the same principle: protect a specific vulnerable area without encapsulating the whole board. Luna: So the decision tree for a startup is: identify the failure mode, then match the material and process to that specific risk. Lucas: Exactly. And I'd add one more thing: test, test, test. Before you commit to potting a whole production run, do a thermal cycle test on potted samples. Run them from minus 40 to plus 85 degrees for a hundred cycles and check for cracks or delamination. A compound that looks great on paper might fail when the board expands and contracts at a different rate than the resin. Luna: Right, because the coefficient of thermal expansion mismatch can cause stress fractures. Lucas: Precisely. The typical CTE for a PCB is around 14 to 17 parts per million per degree Celsius in the plane direction, but through-thickness it's more like 50 to 70. A rigid epoxy might have a CTE of 30 to 40, so it's not a perfect match. If you add a thermally conductive filler, the CTE drops, which can actually help. But you have to validate. Luna: That's a lot of variables. But the payoff — a robust product that survives in the field — is worth the upfront engineering. Lucas: Yeah, and that's the whole point of hardware startups: building something that works reliably, not just a prototype. Potting is one of those unsung manufacturing steps that turns a hobby project into a real product. Luna: Well said. Until next time.